ARTICLE pubs.acs.org/bc
Ultrasmall Gold-Doxorubicin Conjugates Rapidly Kill Apoptosis-Resistant Cancer Cells Xuan Zhang,† Hicham Chibli,† Randall Mielke,‡ and Jay Nadeau*,† † ‡
Department of Biomedical Engineering, McGill University, 3775 University Street, Montreal, Quebec, Canada H3A 2B4 Jet Propulsion Laboratory, California Institute of Technology, Pasadena, California 91109, United States
bS Supporting Information ABSTRACT: Ultrasmall (mean diameter, 2.7 nm) gold nanoparticles conjugated to doxorubicin (Au-Dox) are up to 20-fold more cytotoxic to B16 melanoma cells than the equivalent concentration of doxorubicin alone, and act up to six times more quickly. Ultrasmall Au-Dox enters the cell endocytic vesicles and is also seen free in the cytoplasm and nuclei. This is in distinct contrast to larger particles reported in previous studies, which are excluded from the nucleus and which show no increased toxicity over Dox alone. Cell death with Au-Dox is confirmed to be apoptotic by TUNEL staining and ultrastructural examination using transmission electron microscopy. To further explore the mechanism of action, two other cell lines were examined: HeLa cells which are highly sensitive to Dox, and HeLa cells overexpressing Bcl-2 which show impaired apoptosis and Dox resistance. Interestingly, the Doxsensitive cells show a slightly decreased sensitivity to Au-Dox relative to Dox alone, whereas the Dox-resistant cells are not resistant to AuDox. These results have implications for the design of chemotherapeutic nanoparticles, suggesting that it is possible to selectively target apoptosis-resistant cancer cells while at the same time reducing cytotoxicity to normal cells.
’ INTRODUCTION Doxorubicin (Dox), an anthracycline antibiotic with antineoplastic activity, is among the most widely used anticancer drugs for a wide range of malignancies.1 It is a cytostatic agent with multiple modes of action. Dox can inhibit RNA synthesis by binding to RNA polymerase II and topoisomerase II;2 it may generate reactive oxygen species (ROS) that can cause cell death either by disturbing the cellular redox equilibrium or by damaging the DNA3 or membranes; and it can intercalate into DNA, preventing replication. The major side effects of using doxorubicin in cancer treatment are cell chemoresistance4 and cardiotoxicity.1,5 The type of cardiotoxicity observed is a potentially fatal cardiomyopathy caused by apoptosis of cardiomyocytes, and is an important limiting factor in the clinical use of Dox. Possible methods to improve tumor response and reduce toxicity include the following: (1) improved targeting of Dox to tumor cells; (2) coadministration of drugs that prevent toxicity to normal but not to cancer cells; or (3) modification of Dox to enhance the anticancer mechanisms while reducing the harmful side effects. Nanoparticles have shown preliminary promise for (1). Hydrophilic nanoparticles (NPs) carrying Dox penetrate more deeply into cells than Dox alone, thus reducing the dose needed.4,6-9 Nanoparticles may also be functionalized with a cancer-targeting sequence, such as folic acid, along with the doxorubicin to better target cancer cells.10 Interestingly, nanoparticles improve drug delivery to tumors without specific targeting by the so-called enhanced penetration and retention (EPR) effect, where the vessel leakiness r 2010 American Chemical Society
and impaired lymphatic drainage of cancers causes accumulation of nanometer-sized particles.11 However, point (3) has been addressed poorly, even though it represents a promising avenue for Dox. The mechanisms of Dox action in tumors are different than those seen in cardiomyocytes and endothelial cells, with the activation of apoptotic pathways believed to play a larger role in cancer cells than in normal cells.1 Cancer cells that overexpress antiapoptotic proteins, such as Bcl2, show significant Dox resistance.12,13 A method to overcome these resistance mechanisms while simultaneously delivering the drug to tumors would allow for smaller concentrations of more effective drug to be given to patients. In this work, we find that very small (mean diameter, 2.7 nm) Au nanoparticles conjugated to Dox via an amide bond are taken up into the aggressive, apoptosis-resistant B16 melanoma cell line more efficiently than Dox alone and approximately 6-fold faster (10 min vs 1 h for maximal uptake). Toxicity as measured by the sulforhodamine B (SRB) assay shows that Au-Dox is nearly 20-fold more toxic than Dox alone, showing the greatest effect for Dox concentrations between 0.8 and 10 μM. The particles enter the nucleus as seen by confocal and electron microscopy, in contrast to 5 nm and larger Au particles conjugated to Dox, which show no nuclear entry.14 Interestingly, Received: August 17, 2010 Revised: November 19, 2010 Published: December 28, 2010 235
dx.doi.org/10.1021/bc100374p | Bioconjugate Chem. 2011, 22, 235–243
Bioconjugate Chemistry
ARTICLE
Au-Dox is found to be slightly less toxic than Dox alone to the highly Dox-sensitive HeLa cell line. Stable transfection of HeLa cells with Bcl-2 protects against Dox alone but not against Au-Dox. These results have implications for the design of chemotherapeutic nanoparticles. They indicate that very small particles can enter cell nuclei, and that doxorubicin remains highly active when conjugated to ultrasmall Au, without the need for cleavable linkers to the particle. A great increase in cytotoxicity is seen in apoptosis-resistant cells but not in ordinary, Dox-sensitive cells, suggesting that this conjugate may be useful for chemoresistant cancers.
’ EXPERIMENTAL PROCEDURES Unless otherwise specified, chemicals were purchased from Sigma-Aldrich Canada at the highest grade available and used as delivered. Au-Tiopronin synthesis. The procedure for gold nanoparticle synthesis was adapted from de la Fuente et al.15 Hydrogen tetrachloroaureate(III) trihydrate (0.5 mmol) and tiopronin (N-(2-mercaptopropionyl)glycine) (1.2 mmol) were dissolved in 20 mL of methanol/acetic acid 6:1, and an aqueous solution of sodium borohydrate (7.5 mL, 8 mM) was slowly added. After vigorous stirring for 30 min, the resulting black solution was collected and concentrated. The residues were dissolved in 20 mL H2O and dialyzed for 72 h against dH2O (2 L), which was changed every 12 h. The resulting Au-tiopronin was lyophilized, weighed, and characterized by UV-vis and fluorescence spectroscopy, Fourier transform infrared spectroscopy (FTIR), zeta potential, and transmission electron microscopy (TEM). Particle molecular weight was estimated using the mean diameter found by TEM in order to express particle concentration in moles/liter. UV-vis absorbance spectra were recorded on a Spectra Max Plus plate reader and fluorescence emission spectra on a Gemini EM plate reader (Molecular Devices, Novato, CA). FTIR was performed on a Nicolet FTIR spectrometer (Thermo Scientific). Zeta potential was measured using Zeta plus/Zeta potential analyzer (Brookhaven Instruments Corporation, NY). Samples for TEM were prepared on carbon-coated copper grids, and examined with an accelerating voltage of 200 keV on a JEOL JEM-2100F at the Centre de caracterisation microscopique des cole Polytechnique de Montreal. Also performed materiaux, E were selected area electronic diffraction (SAED) and energy dispersive X-ray spectroscopy (EDS).
Figure 1. Characterization of Au nanoparticles and conjugates. (A) TEM image of Au nanoparticles as synthesized. (B) Size histogram of 1890 particles, fit to a Gaussian with mean size 2.7 nm and standard deviation 0.9 nm. (C) Absorbance (left) and emission (right) spectra of Au nanoparticles as synthesized. No plasmon peak is seen because of the small size of the particles. A weak fluorescence at 780 nm is observed with excitation at 550 nm. (D) Absorbance spectra of 50 μM doxorubicin alone (filled squares), showing quenching when mixed with Au particles (AuþDox) (open squares). This relative absorbance is used to calculate the number of doxorubicin molecules remaining on the dialyzed conjugate (open circles). (E) Schematic of tiopronin-capped Au nanoparticle partially coated with Dox.
Cell Culture. Melanoma and HeLa cells were cultured in high-glucose DMEM (Invitrogen Canada, Burlington, ON) supplemented with L-glutamine (0.2 mM), penicillin (100 U/mL), streptomycin (100 μg/mL), and FBS (10%), and incubated in a 5% CO2 humidified atmosphere. Sulforhodamine Toxicity Assay (SRB). The IC50 of Dox and Au-Dox was determined using the SRB assay. The assay was performed using B16 melanoma cells, HeLa cells stably transfected with the Bcl-2 gene, and HeLa cells stably transfected with a dummy vector. All of these cell lines were a gift of J. Teodoro, McGill Cancer Centre. Cells were passaged at 5 103 cells per well in 96 well culture plates 24 h before use. When they had growth to 60% confluency, they were washed with PBS, incubated with Au alone, Dox alone, or Au-Dox at various concentrations for 30 min in PBS, then washed with PBS and incubated in 200 μL of supplemented DMEM. After 24 or 48 h, cells were fixed with trichloroacetic acid (60 μL of 40% v/v) at 4 °C for 2 h, washed five times with distilled water, air-dried overnight, and stained with SRB reagent (sulforhodamine) (50 μL) for 30 min. Unbound SRB was removed with acetic acid 1%; bound SRB was dissolved in Tris (100 μL of 10 mM solution at pH 10.5).
Au-Tiopronin Conjugation to Doxorubicin and Stability Analysis. The reaction was carried out in 1 mL of 1 PBS
solution containing Au (2 μM), Dox (50 μM), and 1-ethyl-3-(3dimethylaminopropyl)-carbodiimide (EDC 10 μM). Conjugation was performed in the dark at room temperature for 1 h followed by dialysis against dH2O for 6 h using a cellulose dialysis membrane (Sigma) with a MWCO of 12 400. The dialysis membrane was changed after the first hour to prevent a large aggregation of Dox on the membrane. To determine how much free Dox remained in the conjugates, the nanoparticles were removed with an Amicon Ultra-4, MWCO 10 kDa regenerated cellulose membrane and the flowthrough examined with fluorescence spectroscopy (limit of detection, ∼100 nM). For stability analysis, conjugates were incubated at 37 °C either as prepared or with the pH decreased to 5.0 by the addition of concentrated HCl. Absorbance and emission were monitored every hour. 236
dx.doi.org/10.1021/bc100374p |Bioconjugate Chem. 2011, 22, 235–243
Bioconjugate Chemistry
ARTICLE
Figure 2. Microscopic images of labeled B16 cells. (A-D) Confocal microscopy, showing fluorescence in the red channel (excitation 488 nm, emission 590 long-pass) in the left panel and overlay with differential interference contrast (dic) in the right panel. All images taken with the same parameters to allow comparison. (A) Au nanoparticles alone show very little signal. (B) Dox alone shows a classic picture of nuclear staining. (C) Au-Dox showing both nuclear and cytoplasmic fluorescence with overall greater intensity than Dox alone. (D) High-resolution of Dox alone (left) and Au-Dox (right) showing greater overall signal with Au-Dox and significantly increased cytoplasmic staining. (E,F) TEM of uranyl acetate/osmium tetroxide stained thin sections of Au-Dox-labeled cells. The darkest areas are Au (confirmed by energy-dispersive X-ray spectroscopy). (E) Cells in vesicles (left arrow) and the nucleus (right arrow); there is also some Au in the nucleolus (arrowhead). (F) High-resolution image of Au-Dox in and around the nucleus. All of the dark areas are Au. No evident damage to the nuclear membrane can be seen. Note the dispersity of the particles. (G) Unstained thin section showing individual Au particles within an endocytic vesicle.
Absorbance was read at 500 nm. In each plate, at least 5 or 6 repeats were done of each condition, and independent assays were performed at least three times. Fluorescent Labeling and Microscopy. Cells were labeled with fluorescent dyes during the last half hour of incubation with Dox or Au-Dox. When fixed, they were fixed with 2.7% paraformaldehyde for 30 min, washed twice in PBS, and imaged in PBS. When live, they were washed once in PBS and imaged in PBS. Ethidium bromide/acridine orange was used at a final concentration of 11.3 μM. Terminal deoxynucleotidyl transferase dUTP nick-end-labeling (TUNEL) staining was performed using the ApoAlert DNA Fragmentation Assay kit (Clontech). Cells on glass-bottom dishes were treated with 10 μM Dox alone for 24 h (positive control) or 10 μM effective concentration of Au-Dox for 30, 60, or 90 min. Negative controls were untreated or treated with Au only. After treatment, they were fixed with 4% paraformaldehyde for 30 min and washed twice with PBS before permeabilizing with 0.2% Triton-X for 5 min on ice. They were then washed
twice in PBS, incubated in Equilibration Buffer for 10 min, and then labeled with terminal deoxynucleotidyl transferase (TdT) solution for 60 min in the dark at 37 °C. The reaction was stopped with SSC, and the cells were labeled for 5 min with propidium iodide, then washed in H2O and imaged immediately. Wide-field imaging was performed on an Olympus IX-71 with a hyperspectral imaging camera (Chromodynamics, Inc.) that permitted spectral acquisition of output from 400 to 800 nm at increments of g2 nm selected by the user. The excitation filter was a 400-450 bandpass for Dox or a 530-550 bandpass for Au particles. Confocal imaging on live specimens was performed on a Zeiss LSM Pascal with a PlanAchro 40 water immersion objective (NA 0.8). Confocal imaging on fixed specimens was performed on a Zeiss 510 LSM with a PlanApo 100 oil objective. Dyes were excited with an Ar ion laser 488 nm line and/or a green HeNe laser (543 nm line). Acridine orange was observed through a 500-530 nm bandpass filter and ethidium bromide through a 650 nm long-pass filter, a choice which eliminated all channel bleed-through. 237
dx.doi.org/10.1021/bc100374p |Bioconjugate Chem. 2011, 22, 235–243
Bioconjugate Chemistry
ARTICLE
Electron Microscopy of Labeled Cells. Cells for TEM were trypsinized, suspended in PBS, then fixed with 2.5% (v/v) EMgrade glutaraldehyde. After 12-16 h fixation, cells were gently pelleted and washed in H2O. The cells were dispersed into Noble agar worms, stained (or not) with 2% OsO4 and 2% UA, then dehydrated in ethanol and acetone before embedding in Eponate 12 resin. Sample resin blocks were trimmed and sectioned (50-60 nm) on a MT-X Ultramicrotome with a 45° Diatome diamond knife. Ultrathin sections were placed on 200 mesh Formvar/carbon coated copper grids and imaged on a FEI XL 30 with a STEM detector at 30 kV and a working distance of 6.7 mm. Measurement of Reactive Oxygen Species (ROS). ROS was measured using the reporter dye 5-(and-6)-chloromethyl-20 ,70 dichlorodihydrofluorescein diacetate, acetyl ester (CMH2DCFDA) (Invitrogen), which measures ROS accumulation and/or glutathione depletion.16 Cells were incubated with 2 μM Dox or Au-Dox or buffer only, and labeled with the dye according to the manufacturer's instructions at time points of 1 and 3 h. The cells were then trypsinized, washed in PBS, and subjected to flow cytometry on a Beckman Coulter Cell Lab Quanta SC flow cytometer. Excitation was at 488 nm for both the dye and doxorubicin; emission for the dye was collected with a 525 ( 10 nm bandpass filter, and for doxorubicin with a 575þ long-pass filter. At least 25 000 events were collected for each condition. Gating, compensation, and data analysis were performed using FlowJo 9.1 for Macintosh.
’ RESULTS The tiopronin-capped Au nanoparticles showed a relatively broad Gaussian distribution of sizes with mean ( SD of 2.7 ( 0.9 nm (Figure 1A,B). Their zeta potential was -42 mV. Because of their small size, they showed no plasmon absorbance peak. However, they were slightly fluorescent in the far red (peak 780 nm) when excited with green light (550 nm) (Figure 1C). The quantum yield of this fluorescence was 20 cells as a function of time. Data were corrected for photobleaching and background. (B) Images of labeled cells and different time points. Dox or Au-Dox were added at t = 0. The increased background in Au-Dox was typical and represented particles adhering to the culture dish.
The broad orange fluorescence of Dox (emission collected at 590 ( 20 nm) could be used to monitor uptake of Au-Dox and Dox alone in living cells. The Au particles alone were too dimly fluorescent to be observed with confocal microscopy using the 543 nm laser line (Figure 2A) but could be photographed using wide-field epifluorescence with emission at 760-800 nm (not shown). After 1 h of incubation, the overall signal from the AuDox cells was greater than with Dox alone, despite the quenching caused by the Au. The signal was seen in both the cytoplasm and nucleus, in contrast to Dox alone which was almost entirely nuclear (Figure 2B-D). Transmission electron microscopy (TEM) with or without osmium tetroxide/uranyl acetate staining revealed the presence of Au-Dox primarily in endocytic vesicles, but some Au was also observed external to the vesicles and inside the nucleus and nucleolus (Figure 2E,F and Supporting Information, Figure S2). Artifacts of staining made the particles in vesicles appear aggregated; examination of unstained 238
dx.doi.org/10.1021/bc100374p |Bioconjugate Chem. 2011, 22, 235–243
Bioconjugate Chemistry
ARTICLE
Figure 5. Fluorescent TUNEL/propidium iodide staining of B16 cells. Scale bar = 10 μm. (A) Untreated cells showing red labeling throughout the cytoplasm and nucleus. (B) Au-Dox cells after 2 h. 45% of 250 counted cells were positive (green). (C) Higher-magnification image of Au-Dox cells showing a mix of positive and negative cells.
specimens clearly indicated individual Au particles (Figure 2G). Au particles were also endocytosed but not seen in the nucleus (Supporting Information, Figure S2). Dox-only cells at the same time point showed no ultrastructural abnormalities (Supporting Information, Figure S2). The kinetics of Au-Dox entry into cells was very different from that of Dox alone. Time-lapse microscopy with images every 1 min showed a substantial presence of Dox fluorescence in the Au-Dox cells after 10 min, which peaked in 15-20 min. The Dox-only cells showed no measurable fluorescence until 4060 min after drug application (Figure 3). Toxicity independent of kinetics was evaluated with the sulforhodamine B (SRB) assay. Dox, Au-Dox, or control saline was incubated with cells in serum-free medium for a period of time allowing maximal uptake (60-90 min). The cells were then washed and returned to rich medium, and cell density assayed after 24-48 h. B16 cells were relatively resistant to Dox, with an IC50 of ∼17 μM. This fell nearly 20-fold with Au-Dox to ∼930 nM (Figure 4A). The Au particles themselves showed no significant toxicity at up to 30-fold the concentrations used in the conjugates (Figure 4B). In order to determine whether this effect was also seen in cells that were more Dox-sensitive, we repeated the assay with two types of HeLa cells: a cell line stably transfected with a dummy vector containing only a selectable marker (puromycin), and a corresponding cell line stably transfected with Bcl-2, which was
Figure 4. Cytotoxicity of Dox vs Au-Dox measured by the SRB assay. Values are mean ( SEM of 5 duplicate wells in 96-well plates, and lines are fits to the Hill equation with R2 > 0.98 for all. (A) B16 melanoma cells with Au-Dox show significant inhibition relative to Dox alone beginning at concentrations of ∼800 nM. (B) Au nanoparticles only show no toxicity up to 80 μM particles, a concentration more than 30-fold what was used in any of the experiments using conjugates. (C) Untransfected HeLa cells show significant sensitivity to Dox and slightly less sensitivity to Au-Dox. (D) Bcl-2 overexpressing HeLa cells show greatly decreased sensitivity to Dox relative to untransfected cells, but the sensitivity to Au-Dox remains almost unchanged. 239
dx.doi.org/10.1021/bc100374p |Bioconjugate Chem. 2011, 22, 235–243
Bioconjugate Chemistry
ARTICLE
Figure 6. TEM images of B16 cells with (A) Dox alone (20 μM, 24 h incubation) or (B) Au-Dox (10 μM effective Dox concentration, 6 h incubation). Note the large amount of dark staining in the Au-Dox cell, including in the nucleus, which is the Au. The broken chromatin in the nucleus and cellular vacuole formation are characteristic of apoptosis. The lower panels show high-resolution images of the vacuoles, which in some cases are swollen mitochondria (arrow). This image represents a high concentration of Au-Dox in order to clearly show the particles; similar changes were also seen with 1-2 μM Au-Dox (not shown).
expected to be apoptosis-resistant. In the dummy-transfected cells, we found that Au-Dox was slightly less effective than Dox alone, especially at low concentrations. The IC50 of Dox alone was ∼1.5 μM and of Au-Dox was ∼2.5 μM (Figure 4C). However, the picture changed when the cells were stably transfected with the antiapoptotic protein Bcl-2. In this case, the IC50 for Dox rose over 10-fold to >20 μM, but for Au-Dox rose only slightly to 3 μM, a value that was not significantly different from the dummy-vector cells at the 0.95 confidence level (Figure 4D). The mechanism and kinetics of cell death were further investigated using acridine orange/ethidium bromide (AcOr/EtBr) and fluorescent terminal deoxynucleotidyl transferase dUTP nickend-labeling (TUNEL) staining. Low enough concentrations of Dox (50% after 2 h (Figure 5). In addition to these results, TEM examination revealed similar patterns of vacuolization and chromatin condensation typical of apoptosis in both Dox-only and Au-Dox cells (Figure 6).
One of the primary mechanisms of Dox action is generation of reactive oxygen species (ROS),18 especially at high concentrations of drug. At concentrations much higher than that achievable physiologically (the 100 μM range), direct oxidation of membranes has been observed.19 This can also occur with conjugates that adhere to the membrane.20 Overexpression of Bcl-2 can protect against oxidative damage and prevent accumulation of ROS,21 so we used a fluorescent ROS indicator to determine the fraction of ROS-positive cells in untransfected HeLa vs Bcl-2 overexpressing HeLa cells exposed to doxorubicin or Au-Dox. We found that Bcl-2 overexpression prevented any effects of Dox alone at 1 h, and substantially reduced but did not eliminate the effects of Au-Dox. After 1 h, ∼28% of Bcl-2 cells with Au-Dox showed substantial green fluorescence, compared with ∼41% of untransfected HeLa cells with Au-Dox (Figure 7A,B). The cells with Au-Dox clearly showed two distinct populations, so the output in the red channel was also collected to examine the correlation between ROS production and amount of Dox in the cells. We found that the cells showing oxidative effects all contained Dox (Figure 7C,D). As incubation progressed, the ROS signal from Au-Dox labeled cells decreased as the cells died. By 3 h, there was no significant signal from either the HeLa or Bcl-2 cells. However, the signal from Dox-only cells continued to increase in HeLa but not Bcl-2 cells: 16% positive HeLa cells at 3 h, but
